METHOD AND APPARATUS FOR EFFICIENT CODING OF DEPTH LOOKUP TABLE

Description

CROSS REFERENCE TO RELATED APPLICATIONS

[0001] The present invention claims priority to PCT Patent Application, Serial No. PCT/CN2013/070377, filed on January 11, 2013, entitled "An Efficient Coding Method for DLT in 3DVC", PCT Patent Application, Serial No. PCT/CN2013/ 077983, filed on June 26, 2013, entitled "An Efficient Coding Method for DLT in 3DVC" and PCT Patent Application, Serial No. PCT/CN2013/081455, filed on August 14, 2013, entitled "An Efficient Differential Coding Method for DLT in 3DVC".

FIELD OF INVENTION

[0002] The present invention relates to three-dimensional video coding. In particular, the present invention relates to coding the depth lookup table in three-dimensional and multi-view video coding.

BACKGROUND OF THE INVENTION

[0003] Three-dimensional (3D) television has been a technology trend in recent years that intends to bring viewers sensational viewing experience. Various technologies have been developed to enable 3D viewing. Among them, the multi-view video is a key technology for 3D TV application among others. The traditional video is a two-dimensional (2D) medium that only provides viewers a single view of a scene from the perspective of the camera. However, the multi-view video is capable of offering arbitrary viewpoints of dynamic scenes and provides viewers the sensation of realism.

[0004] In 3D video coding, depth information associated with the underlying texture images is useful for improving the coding efficiency as well as for rendering of synthesized views. The statistical distribution of depth data for typical scenes is usually sparse. Some depth values may not occur in depth maps. Accordingly, Simplified Depth Coding (SDC) and Depth Lookup Table (DLT) have been adopted in the HEVC-based Test Model (HTM), where HEVC refers to the emerging video coding standard - High Efficiency Video Coding. In the HTM, for each CU (coding unit), the prediction residual of depth data is usually transform coded. However, the HTM also supports a transform skip mode for the depth data to bypass the transform. When the SDC mode is selected, a set of four different prediction modes is available. After an optimal prediction mode is selected among the four modes, the residual is formed accordingly. In the SDC mode, the transform is not applied to the prediction residual either. Since the SDC prediction process always results in one or two depth segments per coded block, a single residual DC depth value is coded for each of these segments. Furthermore, a DLT is used to map coded depth values in SDC to valid depth values of the original depth map. At the encoder side, the DLT can be constructed based on an initial analysis of the input depth map. The DLT is then coded in the sequence parameter set (SPS).

[0005] The DLT is an optional coding tool for depth map coding. According to the current HTM, the encoder will not use the DLT coding tool if all depth values from 0 to the maximum depth value (e.g. 255) appear in the original depth map during the analysis step. Otherwise, the DLT will be coded in the sequence parameter set (SPS). Fig. 1 illustrates an example of DLT coding process. In order to code the DLT, the number of valid depth values (i.e., 5 in this example) is coded using Exp-Golomb code first. Then, each valid depth value (i.e., a depth value selected from 50, 108, 110, 112 and 200 in this example) is coded using Exp-Golomb code as well. Table 1 illustrates the syntax for DLT coding according to the HTM. Syntax element, dlt_flag[i] indicates whether the DLT coding tool is used for the depth sequence indicated by layer i. If the flag indicates that the DLT coding tool is used (i.e., dlt_flag[i] = 1), syntax element num_depth_values_in_dlt[i] is used to indicate the number of table values in the DLT. The table values in the DLT are then included in the bitstream.

Table 1

dlt_flag[i]	u(1)
if(dlt_flag[i]){
num_depth_values_in_dlt[i]	ue(v)
for (j = 0; j < num_depth_values_in_dlt ; j++){
dlt_depth_value[i][j]	ue(v)
}
}

[0006] Exp-Golomb code is efficient when smaller values have higher probabilities of occurrence. However, that is not the case for valid depth values. As a result, Exp-Golomb codes may not be efficient for DLT coding. In the common test condition, four sequences, i.e., Balloons, Kendo, Newspapercc and PoznanHall are determined to use the DLT coding tool, while the other three are determined not to use the DLT coding tool. Statistics are collected based on the four sequences. As shown in Table 2, DLT requires 557.33 bits in average, which account for more than 65% bits of SPS for depth components. In other words, the DLT coding contributes a majority portion of bits in SPS for depth components. When the SPS carries the DLT, the SPS size becomes much larger than the SPS without the DLT.

Table 2

		DLT Size (bits)	SPS Size (bits)	D/S (%)
Balloons	V0_depth	555	845	65.68%
	V1_depth	641	937	68.41%
	V2_depth	617	915	67.43%
Kendo	V0_depth	485	775	62.58%
	V1_depth	637	933	68.27%
	V2_depth	649	947	68.53%
Newspapercc	V0_depth	606	896	67.63%
	V1_depth	621	917	67.72%
	V2_depth	696	994	70.02%
PoznanHall2	V0_depth	420	712	58.99%
	V1_depth	380	678	56.05%
	V2_depth	381	681	55.95%
Average		557.33	852.50	65.38%

[0007] Accordingly, it is desirable to develop a new DLT coding tool that will improve the coding efficiency of the DLT.

[0008] "3D-CE6.h Results on Simplified Depth Coding with an optional Depth Lookup Table" (FABIAN JAGER, 2012-10-16, 2. JCT-3V MEETING; 102. MPEG MEETING) relates to a modified intra-coding scheme for depth map coding in 3D video. The number of bits to signal the residual values for each segment can be reduced by integrating Depth Lookup Table (DLT), which maps depth values to valid depth values of the original depth map. The DLT is constructed based on an initial analysis of the input depth map and is then coded in the SPS. The analysis of the 3DV test sequences shown that the estimated depth maps do not utilize the full available depth range of 2⁸. The parameter num_depth_values_in_dlt specifies the number of different depth values and the number of elements in the depth lookup table. The parameter dlc_depth_value[i] specifies a valid depth value and an entry in the depth lookup table.

[0009] "3D-HEVC Test Model 2" (TECH G ET AL, 2012-12-23, 2. JCT-3V MEETING; 102, MPEG MEETING) relates to draft of 3D-HEVC Test Model Description. The Depth Lookup Table utilizes the property of the depth map, that the full available depth range of 28 values is not utilized. Only a small amount of different depth levels occur due to strong quantization. This depth lockup-table is used during the coding process to reduce the effective signal bit-depth of the residual signal. Video parameter set extension syntax and decoding process for a depth lookup table are discussed.

[0010] A Beginners Guide to Bitmaps" (Paul Bourke, 1993-11-01) relates to an elementary introduction to bitmaps as they are used in computer graphics. Bitmaps are defined as a regular rectangular mesh of cells called pixels, each pixel containing a colour value. They are characterised by only two parameters, the number of pixels and the information content (colour depth) per pixel. There are other attributes that are applied to bitmaps but they are derivations of these two fundamental parameters.

SUMMARY OF THE INVENTION

[0011] The above-mentioned objectives are achieved by a method for decoding according to independent claim 1 and by a method for encoding according to independent claim 7. The dependent claims define preferred embodiments thereof.

[0012] A method and apparatus for encoding and decoding depth lookup table (DLT) in a three-dimensional or multi-view video coding using a constrained table value range are disclosed. For a decoder, embodiments according to the present invention decode a constrained table value range from a minimum table value to a maximum table value of the DLT based on the coded DLT information. The minimum table value is greater than a minimum depth value of a full-range DLT, the maximum table value is smaller than a maximum depth value of the full-range DLT, or the minimum table value is greater than the minimum depth value and the maximum table value is smaller than the maximum depth value. Bit map values for depth values in a depth range from (the minimum table value + 1) to (the maximum table value - 1) are then decoded. The bit map values indicate whether the depth values are in the DLT respectively. The DLT for the current depth map can be reconstructed based on the constrained table value range and the bit map values for the depth values in the depth range. For an encoder, embodiments according to the present invention determine a constrained table value range from a minimum table value to a maximum table value of the DLT and encode the DLT into coded DLT information, wherein said encoding the DLT into the coded DLT information comprises encoding first information associated with the constrained table value range into the coded DLT information.

[0013] A flag can be used to indicate whether to use the full-range DLT or a constrained DLT. The constrained table value range may be represented by the minimum table value and a maximum table difference, wherein the maximum table difference corresponds to the difference between the maximum table value and the minimum table value, and the minimum table value and the maximum table difference can be coded using unsigned integers. The minimum table value and the maximum table value may use default values when the minimum table value and the maximum table difference are not coded. For example, the minimum table value may use -1 as a default value and the maximum table value may use MAX_DEPTH_VALUE + 1 as the default value, wherein MAX_DEPTH_VALUE corresponds to the maximum depth value. The constrained table value range may also be represented by the minimum table value and the maximum table value, and the minimum table value and the maximum table value can be coded using unsigned integers.

[0014] Another aspect of the present invention addresses bit map coding. The bit map values for depth values in a depth range from (the minimum table value + 1) to (the maximum table value - 1) are also incorporated in the coded DLT information, where the bit map values indicate whether respective depth values are in the DLT or not. In one embodiment, the bit map values for the depth values in the depth range are coded using run-length coding on zero-runs or one-runs associated with the bit map values. In another embodiment, the bit map values for the depth values in the depth range are coded using differences between the zero-runs and a minimum zero-run or between the one-runs and a minimum one -run. The differences can be encoded into the coded DLT information using n-bit fixed length coding, where n is a positive integer. The excess part of the second difference can be coded using Exp-Golomb. An optimal value for n can be determined at the encoder side based on a lowest resulting bitrate for the coded DLT information. A flag can be coded to signal whether the run-length coding is used for coding bit map values or the bit map values are coded as a bit-map directly.

[0015] In yet another embodiment of the present invention, the bit map values for the depth values in the depth range are coded using inter-view coding when the current depth map is in a dependent view. The inter-view bit map values are formed by applying Exclusive-OR operations between the depth values in the depth range and second bit map values associated with a second DLT of a second depth map in a base view.

BRIEF DESCRIPTION OF THE DRAWINGS

[0016] 

Fig. 1 illustrates an example of a depth lookup table (DLT) consisting of table values 50, 108, 110, 112 and 200.

Fig. 2 illustrates an example of depth lookup table (DLT) coding according to an embodiment of the present invention, where a full map flag, the minimum table value, the maximum table value and the bit map for table values between the minimum table value and the maximum table value are coded.

Fig. 3 illustrates an example of depth lookup table (DLT) coding according to an embodiment of the present invention, where the bit map is coded using run-length coding.

Fig. 4 illustrates an example of depth lookup table (DLT) coding according to an embodiment of the present invention, where the inter-view bit map is derived based on the bit map of the base view.

Fig. 5 illustrates an example of depth lookup table (DLT) coding according to an embodiment of the present invention, where the differences of consecutive table values are coded.

Fig. 6 illustrates an exemplary flow chart for decoding a depth lookup table according to an embodiment of the present invention.

Fig. 7 illustrates an exemplary flow chart for encoding a depth lookup table according to an embodiment of the present invention.

DETAILED DESCRIPTION

[0017] Reference throughout this specification to "one embodiment," "an embodiment," or similar language means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present invention. Thus, appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment.

[0018] The illustrated embodiments of the invention will be best understood by reference to the drawings, wherein like parts are designated by like numerals throughout. The following description is intended only by way of example, and simply illustrates certain selected embodiments of apparatus and methods that are consistent with the invention as claimed herein.

[0019] The present invention discloses a coding method that indicates the range of the DLT for efficient DLT coding. In one embodiment, a Range Constrained Bit Map (RCBM) coding method is disclosed. According to RCBM, the DLT is coded in two steps.

[0020] In the first step, the minimum table value (i.e., min_table_value) and the maximum table value difference (i.e., diff_max_value) are coded as unsigned integers to indicate the constrained value range of the DLT. The largest table value is designated as max_table_value, which can be determined according to max_table_value = min_table_value + diff max_value.

[0021] In the second step, the bit map is coded, where a flag, bit_map_flag[i] is used to indicate whether an underlying depth value exists for the DLT to be coded. Since both max_table _value and min_table_value exist, there is no need of bit map for max_table_value and min_table_value. Accordingly the index i for bit_map_flag[i] is from 0 to (max_table_value - min_table_value - 2). If bit_map_flag [i] =1, then depth value (i + min_table_value + 1) of the DLT exists. If bit_map_flag [i] =0, depth value (i + min_table_value + 1) of the DLT does not exists.

[0022] A flag, full_map_flag is used to signal whether to use a DLT with a constrained table value range, where the constrained table value range is identified by min table_value_and diff_max_value. When the range is not coded, default values can be used. For example, min_table_value can use a default value of -1, max_table_value can use a default value of (MAX_DEPTH_VALUE + 1). A full bit map without range constrained is coded in this case. The full bit map requires (MAX_DEPTH_VALUE + 1) bits. The encoder can decide whether to code the DLT table according to the RCBM depending on whether RCBM results in fewer bits. According to this embodiment, the maximum number of bits required to code the DLT is MAX_DEPTH_VALUE + 2 (e.g. 257 for 8-bit samples).

[0023] Fig. 2 illustrates exemplary coding process based on the RCBM for the DLT in Fig. 1. In Fig. 2, full_map_flag = 0 indicates that the full map is not used as shown in step 210. The minimum table value, i.e., min_table_value = 50 is coded in step 220. The table range, i.e., diff_max_value = 150 is coded in step 230. The bit map values, i.e., bit_map_flag[i] for i = 0,..., 148 (i.e., 149 bits) are coded as shown in step 240.

[0024] Exemplary syntax to support the RCBM is shown in Table 3. Syntax element, code_full_bit_map_[i] is a flag used to indicate whether a full map is used or not. When the full bit map is not used, the minimum table value (i.e., min_dlt_value[i]) and the range value (i.e., diff_max_dlt_value[i]) are incorporated in the syntax. If the full map is used, the default min_dlt_value[i] and diff_max_dlt_value[i] (i.e., -1 and MAX_DEPTH_VALUE + 1) will be used respectively. The bit map, i.e., bit_map_flag[i][j] is then coded.

Table 3

code full_bit_map_ftag[i]	u(1)
if(!code_full_bit_map_flag[i]){
min_dlt_value[i]	u(v)
diff_max_dlt_value[i]	u(v)
}
for(j=0;j< max_dlt_value[i] - min_dlt_value[i] - 1;j++)
bit_map_flag[i][j]	u(1)

[0025] In another embodiment of the present invention, the DLT is coded using an enhanced range constrained bit-map (ERCBM) coding method. According to the ERCBM, the DLT is represented by a range constrained bit-map as mentioned above, where min_dlt_value and diff_ max_dlt_value are coded as unsigned integers to constrain the table value to the range [MinDltValue, MaxDltValue] for the DLT. Values outside this range are not used in the DLT. Values in DLT between (MinDltValue, MaxDltValue) can be represented by a bit-map BM[i] with i from 0 to MaxDltValue - MinDltValue - 2. If BM[i] = 1, then value (i + MinDltValue + 1) belongs to the DLT. Otherwise, value (i + MinDltValue + 1) does not belong to DLT.

[0026] In some embodiments, differential run-length coding is applied to code the bit-map. The differential run-length coding is also referred to as run-length coding or differential coding in the following descriptions. The bit-map can also be represented by the lengths of zero-runs (i.e., the number of consecutive zeroes) or one-runs (i.e., the number of consecutive ones). Run-length coding is efficient for underlying data having high probabilities of repeating symbols (i.e., 0 in this case). For the bit map of the DLT with a small number of table values, run length coding on zero-runs will be more efficient. Fig. 3 illustrates exemplary differential run-length coding incorporating an embodiment of the present invention. The lengths of zero-run are 3, 2, 2 and 3 sequentially for the DLT in Fig. 3, wherein the first zero-run (i.e., 3) corresponds to three 0's between depth value 64 and depth value 60; the second zero-run corresponds to the two 0's between depth value 67 and depth value 64; the third zero-run corresponds to the two 0's between depth value 70 and depth value 67; and the fourth zero-run corresponds to the three 0's between depth value 74 and depth value 70. The zero-runs correspond to 3, 2, 2 and 3 respectively are coded along with the minimum table value and the range value (i.e. the difference between the maximum table value and the minimum table value). The zero-runs can be coded by fixed-length coding with n bits. If a difference exceeds 2^n-2, the excess will be coded with Exp-Golomb code, where n is signaled before the differences are coded. The encoder can choose an optimal n that may result in the minimum number of bits.

[0027] In order to guarantee the number of bits for the worst case to be predictable and acceptable, the encoder can use the RCBM as a fall back mode. In the worst case, 0, 1, ... 255 are included in the DLT. diff value requires 8 bits, min_value requires 0 bit, run length coding_flag requires 1 bit, and the bit_map_flags requires 254 bits when RCBM is utilized. Accordingly, a DLT can be coded with 263 bits in the worst case for 8-bit depth data. An exemplary syntax design according to an embodiment of the present invention are illustrated in Table 4, where min_run_length corresponds to the minimum zero-run, run_length_bits_minus1 corresponds to (n-1) and run_length_diff[j] corresponds to the difference between a current zero-run and the minimum zero-run.

Table 4

diff_value	u(v)
min_value	u(v)
run_length _coding_flag	u(1)
if(run_length_coding_flag){
min_run_length	u(3)
run_length_bits_minus1	u(3)
for(j = MinV+1; j< MaxV; j+=Run[j]+1){
run_length_diff[j]	u(v)
if(run_length_diff[j]==RunBoundary)
run_length_diff_rem[j]	ue(v)
}
}
else{
for(j = MinV + 1; j< MaxV; j++)
bit_map_flag[j]	u(1)
}

[0028] The differential run-length coding for the DLT as disclosed above can be combined with other DLT processing techniques. For example, the differential run-length coding can be used jointly with the delta-DLT method as disclosed in JCT3V-D0054 (Fabian Jäger, "CE6.H related: Results on Updating Mechanism for Coding of Depth Lookup Table (Delta-DLT)", Joint Collaborative Team on 3D Video Coding Extensions of ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 1, 4th Meeting: Incheon, KR, 20-26 Apr. 2013, Document: JCT3V-D0054). In JCT3V-D0054, the inter-view correlation is utilized to improve the coding efficiency. The delta-DLT process is illustrated in Fig. 4, where DLT0 and DLT1 correspond to the DLTs of the base view and the dependent view respectively. Both DLT0 and DLT1 are represented in the full bit-map form. DLT1' is calculated as DLT1' = DLT0⊕DLT1, where "⊕" represents the "Exclusive OR" operation. According to the delta-DLT method, the DLT1' is coded instead of DLT1. After DLT1' is decoded, DLT1 is reconstructed as DLT1 = DLT0⊕DLT1'. Since DLT0 and DLT1 correspond to the bit maps in two views of a same scene, they should be very similar. Therefore, the Exclusive OR operation will result a lot of zeroes, which can be beneficial for length coding of zero-runs. Since the inter-view information (i.e., DLTO) is only utilized at the reconstruction step, the delta-DLT approach does not introduce any parsing dependency issue. Therefore, ERCBM can work well regardless DLTs are coded in VPS (video parameter set), SPS (sequence parameter set), PPS (picture parameter set) or slice headers.

[0029] A differential coding method is disclosed where min_dlt_value and diff_max_dlt_value are coded as unsigned integers to indicate the constrained table value range, [MinDltValue, MaxDltValue] of the DLT as before. Values outside this range are not used in the DLT. Differences between consecutive values in DLT can also be used to represent the DLT. Difference, i.e., Diff[i] is defined as (DLTValue[i+1] - DLTVlaue[i] - 1). For example, if table values in the DLT are 60, 64, 67, 70 and 74, the corresponding differences are 3, 2, 2 and 3. The differences between consecutive values can also be interpreted as the zero-runs.

[0030] In order to further improve differential or run-length coding efficiency, the minimum difference (i.e. the minimum zero-run) is coded first and then deltas between respective differences (i.e. zero-runs) and the minimum difference are coded subsequently. Fig. 5 demonstrates an example of a DLT, where differences are 3, 2, 2 and 3. The minimum difference '2' is coded first, and then the deltas corresponding to the differences, i.e., 1, 0, 0 and 1 are coded in order. The deltas are coded by fixed-length coding with n bits. If a difference exceeds 2ⁿ-2, the excess will be coded with Exp-Golomb code. The number of bits for the fixed-length coding, i.e., n is signaled before deltas are coded. Since n is always greater then 0, n-1 instead of n is coded. The encoder can choose the optimal n that results in the minimum number of bits for the DLT.

[0031] The coding process using the delta of the difference between an underlying value and the minimum value in the DLT can be described based on Fig. 5. At the encoder side, the encoding process is as follows:

a. The difference between the minimum value '60' and maximum value '74' is '14'. Therefore, diff_max_dlt_value =14 is coded first.

b. min_dlt_value = 60 is then coded.

c. min_diff =2 is coded.

d. Deltas associated with the table values in the DLT correspond to 1, 0, 0 and 1. If n = 1, the required bits for the deltas are 2, 1, 1 and 2 respectively, resulting in a total of 6 bits. If n = 2, the required bits for the deltas will be 2, 2, 2 and 2 respectively, resulting in a total of 8 bits. If n is larger, the total bits will be even more. Consequently, the encoder will choose n = 1. Thus, diff_bits_minus1 will have a value of (n-1) = 0.

e. DiffBoundary, which is defined as 2ⁿ-1, is equal to 1. Syntax elements diff_minus_min (i.e., the delta between a current difference and the minimum difference) and diff_minus_min_rem (i.e., the excess part of the delta) are coded as:

diff_minus_min[0]=1, diff_minus_min_rem[0]=0;

diff_minus_min[1]=0, diff_minus_min_rem[1] not coded;

diff_minus_min[2]=0, diff_minus_min_rem[2] not coded;

diff_minus_min[3]=1, diff_minus_min_rem[3]=0;

[0032] At the decoder side, the decoding process is as follows:

a. DLTValue[0] = 60 since min_dlt_value = 60. Maximum value in DLT is 60 + 14 = 74;

b. Diff[0] = 2+1+0=3, so DLTValue[1] = 60+ 3 + 1 = 64;

c. Diff[1] = 2+0=2, so DLTValue[2] = 64+ 2 + 1 = 67;

d. Diff[2] = 2+0=3, so DLTValue[3] = 67+ 2 + 1 = 70;

e. Diff[3] = 2+1+0=3, so DLTValue[5] = 70+ 3 + 1 = 74; 74 is equal to the maximum value in DLT, so stop parsing differences.

[0033] Exemplary syntax incorporating deltas of differences based on an embodiment of the present invention is shown in Table 5. Flag DLTDeduced is used to indicate whether the DLT can be deduced from diff_max_dlt_value and min_dlt_value. If diff_max_dlt_value is 0, there will be only one value in the DLT (i.e., min_dlt_value) and the flag DLTDeduced is set to 1. Furthermore, if diff_max_dlt_value is 1, there will be only two values in the DLT (i.e., min_dlt_value and min_dlt_value+1), DLTDeduced is also set to 1. In all other cases, DLTDeduced is set to 0.

Table 5

diff_max_dlt_value	u(v)
min_dlt_value	u(v)
if(!DLTDeduced){
min_diff	u(3)
diff_bits_minus1	u(3)
for(j =0; DLTValue[j]<MaxDLTValue; j++){
diff_minus_min[j]	u(v)
if(diff_minus_min[j]==DiffBoundary)
diff_minus_min_rem[j]	ue(v)
}
}

[0034] In one embodiment, a flag is coded to signal whether the differential run-length coding is used for coding bit map values of the DLT. For example, the encoder and the decoder apply the bit map coding method to code the bit map values of the DLT as a bit-map directly when the flag is 0, and apply the differential run-length coding method to code the bit map values of the DLT when the flag is 1. The bit map values indicate whether the depth values are in the DLT respectively.

[0035] Fig. 6 illustrates an exemplary flowchart for depth lookup table (DLT) decoding in a three-dimensional or multi-view video coding system incorporating a constrained table value range according to an embodiment of the present invention. The system receives input data associated with coded DLT information of a current depth map as shown in step 610. The input data may also be received from a processor such as a controller, a central processing unit, a digital signal processor or electronic circuits that produce the input data. A constrained table value range from a minimum table value to a maximum table value of the DLT is decoded based on the coded DLT information as shown in step 620. The minimum table value is greater than a minimum depth value of a full-range DLT, the maximum table value is smaller than a maximum depth value of the full-range DLT, or the minimum table value is greater than the minimum depth value and the maximum table value is smaller than the maximum depth value. Bit map values for depth values in a depth range from (the minimum table value + 1) to (the maximum table value - 1) are then decoded as shown in step 630, where the bit map values indicate whether the depth values are in the DLT respectively. The DLT for the current depth map is then reconstructed based on the constrained table value range and the bit map values for the depth values in the depth range.

[0036] Fig. 7 illustrates an exemplary flowchart for depth lookup table (DLT) encoding in a three-dimensional or multi-view video coding system incorporating a constrained table range according to an embodiment of the present invention. The system receives input data associated with the DLT of a current depth map as shown in step 710. The input data may also be received from a processor such as a controller, a central processing unit, a digital signal processor or electronic circuits that produce the input data. A constrained table value range from a minimum table value to a maximum table value of the DLT is determined as shown in step 720. The minimum table value is greater than a minimum depth value of a full-range DLT, the maximum table value is smaller than a maximum depth value of the full-range DLT, or the minimum table value is greater than the minimum depth value and the maximum table value is smaller than the maximum depth value. The DLT is then encoded into coded DLT information as shown in step 730, wherein said encoding the DLT into the coded DLT information comprises encoding first information associated with the constrained table value range into the coded DLT information.

[0037] The flowcharts shown above are intended to illustrate examples of depth lookup table with a constrained table range.

[0038] The above description is presented to enable a person of ordinary skill in the art to practice the present invention as provided in the context of a particular application and its requirement. Various modifications to the described embodiments will be apparent to those with skill in the art, and the general principles defined herein may be applied to other embodiments. Therefore, the present invention is not intended to be limited to the particular embodiments shown and described, but is to be accorded the widest scope consistent with the principles and novel features herein disclosed. In the above detailed description, various specific details are illustrated in order to provide a thorough understanding of the present invention. Nevertheless, it will be understood by those skilled in the art that the present invention may be practiced.

[0039] Embodiment of the present invention as described above may be implemented in various hardware, software codes, or a combination of both. For example, an embodiment of the present invention can be a circuit integrated into a video compression chip or program code integrated into video compression software to perform the processing described herein. An embodiment of the present invention may also be program code to be executed on a Digital Signal Processor (DSP) to perform the processing described herein. The invention may also involve a number of functions to be performed by a computer processor, a digital signal processor, a microprocessor, or field programmable gate array (FPGA). These processors can be configured to perform particular tasks according to the invention, by executing machine-readable software code or firmware code that defines the particular methods embodied by the invention. The software code or firmware code may be developed in different programming languages and different formats or styles. The software code may also be compiled for different target platforms. However, different code formats, styles and languages of software codes and other means of configuring code to perform the tasks in accordance with the invention will not depart from the spirit and scope of the invention.

Claims

1. A method for decoding a depth lookup table, DLT, in a three-dimensional or multi-view video coding system, the method comprising:

receiving input data associated with coded DLT information of a current depth map;

decoding a constrained table-value range for the DLT from a minimum table value (Vmin) of the DLT to a maximum table value (Vmax) of the DLT based on the coded DLT information, and wherein the minimum table value (Vmin) is greater than a minimum depth value of a full-range DLT, the maximum table value (Vmax) is smaller than a maximum depth value of the full-range DLT, or the minimum table value (Vmin) is greater than the minimum depth value and the maximum table value (Vmax) is smaller than the maximum depth value;

decoding bit map values of a bit map (BM) for depth values in a depth range from (the minimum table value + 1) to (the maximum table value - 1), wherein the bit map values indicate whether the depth values are present in the DLT respectively; and

reconstructing the DLT for the current depth map based on the constrained table-value range for the DLT and the bit map values for the depth values in the depth range;

wherein the constrained table-value range for the DLT is indicated by the minimum table value (Vmin) and a maximum table value difference in the coded DLT information, wherein the maximum table value difference corresponds to a first difference between the maximum table value (Vmax) and the minimum table value (Vmin), and wherein the minimum table value (Vmin) and the maximum table value difference are coded using unsigned integers; or

wherein the constrained table-value range for the DLT is indicated by the minimum table value (Vmin) and the maximum table value (Vmax) coded in the coded DLT information, wherein the minimum table value (Vmin) and the maximum table value (Vmax) are coded using unsigned integers.

2. The method of claim 1, further comprising decoding a first flag to indicate whether to use the full-range DLT or a constrained DLT, wherein the constrained DLT has the constrained table-value range for the DLT from the minimum table value (Vmin) to the maximum table value (Vmax).

3. The method of Claim 1, wherein the minimum table value (Vmin) and the maximum table value (Vmax) use default values when the minimum table value (Vmin) and the maximum table value difference are not coded in the coded DLT information and wherein the minimum table value (Vmin) uses a first default value corresponding to - 1 and the maximum table value (Vmax) uses a second default value corresponding to MAX_DEPTH_VALUE + 1, wherein MAX_DEPTH_VALUE corresponds to the maximum depth value.

4. The method of any of the preceding claims, wherein the bit map values for the depth values in the depth range are coded using run-length coding on zero-runs or one-runs associated with the bit map values.

5. The method of Claim 4, wherein second differences between the zero-runs and a minimum zero-run or between the one-runs and a minimum one-run are coded, said method further comprising
decoding a parameter n from the coded DLT information, wherein the second differences are coded using n-bit fixed length coding and n is a positive integer, and
decoding an excess part of a current second difference from the coded DLT information when said n-bit fixed length coding indicates that the current second difference exceeds a maximum n-bit value represented by n bits, wherein the excess part is coded using an Exp-Golomb code, and wherein the excess part corresponds to a third difference between the current second difference and the maximum n-bit value.

6. The method of Claim 4, wherein a flag is signaled to indicate whether the DLT is coded using the run-length coding or the DLT is coded as a bit-map directly.

7. A method for encoding a depth lockup table, DLT, in a three-dimensional or multi-view video coding system, the method comprising:

receiving input data associated with the DLT of a current depth map;

determining a constrained table-value range for the DLT from a minimum table value (Vmin) to a maximum table value (Vmax) of the DLT, wherein the minimum table value (Vmin) is greater than a minimum depth value of a full-range DLT, the maximum table value (Vmax) is smaller than a maximum depth value of the full-range DLT, or the minimum table value (Vmin) is greater than the minimum depth value and the maximum table value (Vmax) is smaller than the maximum depth value;

encoding bit map values of a bit map (BM) for depth values in a depth range from (the minimum table value + 1) to (the maximum table value - 1), wherein the bit map values indicate whether the depth values are present in the DLT respectively; and

encoding the DLT into coded DLT information, wherein said encoding the DLT into the coded DLT information comprises encoding first information associated with the constrained table-value range for the DLT into the coded DLT information;

wherein said first information associated with the constrained-table value range for the DLT consists of the minimum table value (Vmin) and a maximum table value difference, wherein the maximum table value difference corresponds to a first difference between the maximum table value (Vmax) and the minimum table value (Vmin), and wherein the minimum table value (Vmin) and the maximum table value difference are coded using unsigned integers; or

wherein said first information associated with the constrained-table value range for the DLT consists of the minimum table value (Vmin) and the maximum table value (Vmax), wherein the minimum table value (Vmin) and the maximum table value (Vmax) are coded using unsigned integers.

8. The method of claim 7, further comprising encoding a first flag to indicate whether to use the full-range DLT or a constrained DLT, wherein the constrained DLT has the constrained table-value range for the DLT from the minimum table value (Vmin) to the maximum table value (Vmax).

9. The method of claim 7, wherein the minimum table value (Vmin) and the maximum table value (Vmax) use default values when the minimum table value (Vmin) and the maximum table value difference are not coded, and
wherein the minimum table value (Vmin) uses a first default value corresponding to - 1 and the maximum table value (Vmax) uses a second default value corresponding to MAX_DEPTH_VALUE + 1, wherein MAX_DEPTH_VALUE corresponds to the maximum depth value.

10. The method of claim 7, wherein the bit map values for the depth values in the depth range are coded using run-length coding on zero-runs or one-runs associated with the bit map values.

11. The method of claim 10, wherein second differences between the zero-runs and a minimum zero-run or between the one-runs and a minimum one-run are encoded.

12. The method of claim 11, further comprising encoding a flag to indicate whether the DLT is encoded using the run-length coding or the DLT is encoded as a bit-map directly.

13. The method of claim 11, wherein the second differences are encoded into the coded DLT information using n-bit fixed length coding, wherein n is a positive integer.

14. The method of claim 13, wherein an excess part of a current second difference is coded using an Exp-Golomb code when the current second difference exceeds a maximum n-bit value represented by n bits, wherein the excess part corresponds to a third difference between the current second difference and the maximum n-bit value.

Ansprüche

1. Verfahren zum Decodieren einer Tiefen-Nachschlagtabelle, DLT, in einem System zum Codieren von dreidimensionalem oder Mehrfachansicht-Video, wobei das Verfahren aufweist:

Empfangen von Eingangsdaten, die zu codierten DLT-Informationen einer aktuellen Tiefenzuordnung gehören;

Decodieren eines eingeschränkten Tabellenwertebereichs für die DLT von einem minimalen Tabellenwert (Vmin) der DLT bis zu einem maximalen Tabellenwert (Vmax) der DLT basierend auf den codierten DLT-Informationen, und wobei der minimale Tabellenwert (Vmin) größer ist als ein minimaler Tiefenwert einer Vollbereichs-DLT, der maximale Tabellenwert (Vmax) kleiner ist als ein maximaler Tiefenwert der Vollbereichs-DLT, oder der minimale Tabellenwert (Vmin) größer ist als der minimale Tiefenwert und der maximale Tabellenwert (Vmax) kleiner ist als der maximale Tiefenwert;

Decodieren von Bit-Zuordnungswerten einer Bit-Zuordnung (BM) für Tiefenwerte in einem Tiefenbereich von (dem minimalen Tabellenwert + 1) bis zu (dem maximalen Tabellenwert - 1), wobei die Bit-Zuordnungswerte anzeigen, ob die Tiefenwerte jeweils in der DLT vorhanden sind; und

Rekonstruieren der DLT für die aktuelle Tiefenzuordnung basierend auf dem eingeschränkten Tabellenwertebereich für die DLT und den Bit-Zuordnungswerten für die Tiefenwerte in dem Tiefenbereich;

wobei der eingeschränkte Tabellenwertebereich für die DLT durch den minimalen Tabellenwert (Vmin) und eine maximale Tabellenwertdifferenz in den codierten DLT-Informationen angezeigt wird, wobei die maximale Tabellenwertdifferenz zu einer ersten Differenz zwischen dem maximalen Tabellenwert (Vmax) und dem minimalen Tabellenwert (Vmin) korrespondiert, und wobei der minimale Tabellenwert (Vmin) und die maximale Tabellenwertdifferenz unter Verwendung von vorzeichenlosen ganzen Zahlen codiert werden; oder

wobei der eingeschränkte Tabellenwertebereich für die DLT durch den minimalen Tabellenwert (Vmin) und den maximalen Tabellenwert (Vmax), die in den codierten DLT-Informationen codiert sind, angezeigt wird, wobei der minimale Tabellenwert (Vmin) und der maximale Tabellenwert (Vmax) unter Verwendung von vorzeichenlosen ganzen Zahlen codiert werden.

2. Verfahren gemäß Anspruch 1, das weiter ein Decodieren eines ersten Flags aufweist, das anzeigt, ob die Vollbereichs-DLT oder eine eingeschränkte DLT zu verwenden ist, wobei die eingeschränkte DLT den eingeschränkten Tabellenwertebereich für die DLT von dem minimalen Tabellenwert (Vmin) bis zu dem maximalen Tabellenwert (Vmax) aufweist.

3. Verfahren gemäß Anspruch 1, wobei der minimale Tabellenwert (Vmin) und der maximale Tabellenwert (Vmax) Vorgabewerte verwenden, wenn der minimale Tabellenwert (Vmin) und die maximale Tabellenwertdifferenz nicht in den codierten DLT-Informationen codiert sind, und wobei der minimale Tabellenwert (Vmin) einen ersten Vorgabewert verwendet, der zu -1 korrespondiert, und der maximale Tabellenwert (Vmax) einen zweiten Vorgabewert verwendet, der zu MAX_DEPTH_VALUE + 1 korrespondiert, wobei MAX_DEPTH_VALUE zu dem maximalen Tiefenwert korrespondiert.

4. Verfahren gemäß einem der vorstehenden Ansprüche, wobei die Bit-Zuordnungswerte für die Tiefenwerte in dem Tiefenbereich unter Verwendung von Lauflängen-Codierung auf Null-Läufen oder Eins-Läufen codiert werden, die zu den Bit-Zuordnungswerten gehören.

5. Verfahren gemäß Anspruch 4, wobei zweite Differenzen zwischen den Null-Läufen und einem minimalen Null-Lauf oder zwischen den Eins-Läufen und einem minimalen Eins-Lauf codiert werden, wobei das Verfahren weiter aufweist
Decodieren eines Parameters n aus den codierten DLT-Informationen, wobei die zweiten Differenzen unter Verwendung von n-Bit-Festlängen-Codierung codiert werden und n eine positive ganze Zahl ist, und
Decodieren eines überschüssigen Teils einer aktuellen zweiten Differenz aus den codierten DLT-Informationen, wenn die n-Bit-Festlängen-Codierung anzeigt, dass die aktuelle zweite Differenz einen maximalen n-Bit-Wert überschreitet, der durch n Bits repräsentiert wird, wobei der überschüssige Teil unter Verwendung eines Exp-Golomb-Codes codiert wird und wobei der überschüssige Teil zu einer dritten Differenz zwischen der aktuellen zweiten Differenz und dem maximalen n-Bit-Wert korrespondiert.

6. Verfahren gemäß Anspruch 4, wobei ein Flags signalisiert wird, das anzeigt, ob die DLT unter Verwendung der Lauflängen-Codierung codiert ist oder die DLT direkt als eine Bit-Zuordnung codiert ist.

7. Verfahren zum Codieren einer Tiefen-Nachschlagtabelle, DLT, in einem System zum Codieren von dreidimensionalem oder Mehrfachansicht-Video, wobei das Verfahren aufweist:

Empfangen von Eingangsdaten, die zu der DLT einer aktuellen Tiefenzuordnung gehören;

Bestimmen eines eingeschränkten Tabellenwertebereichs für die DLT von einem minimalen Tabellenwert (Vmin) bis zu einem maximalen Tabellenwert (Vmax) der DLT, wobei der minimale Tabellenwert (Vmin) größer ist als ein minimaler Tiefenwert einer Vollbereichs-DLT, der maximale Tabellenwert (Vmax) kleiner ist als ein maximaler Tiefenwert der Vollbereichs-DLT, oder der minimale Tabellenwert (Vmin) größer ist als der minimale Tiefenwert und der maximale Tabellenwert (Vmax) kleiner ist als der maximale Tiefenwert;

Codieren von Bit-Zuordnungswerten einer Bit-Zuordnung (BM) für Tiefenwerte in einem Tiefenbereich von (dem minimalen Tabellenwert + 1) bis zu (dem maximalen Tabellenwert - 1), wobei die Bit-Zuordnungswerte anzeigen, ob die Tiefenwerte jeweils in der DLT vorhanden sind; und

Codieren der DLT in codierte DLT-Informationen, wobei das Codieren der DLT in die codierten DLT-Informationen ein Codieren von ersten Informationen, die zu dem eingeschränkten Tabellenwertebereich für die DLT gehören, in die codierten DLT-Informationen umfasst;

wobei die ersten Informationen, die zu dem eingeschränkten Tabellenwertebereich für die DLT gehören, aus dem minimalen Tabellenwert (Vmin) und einer maximalen Tabellenwertdifferenz bestehen, wobei die maximale Tabellenwertdifferenz zu einer ersten Differenz zwischen dem maximalen Tabellenwert (Vmax) und dem minimalen Tabellenwert (Vmin) korrespondiert, und wobei der minimale Tabellenwert (Vmin) und die maximale Tabellenwertdifferenz unter Verwendung von vorzeichenlosen ganzen Zahlen codiert werden; oder

wobei die ersten Informationen, die zu dem eingeschränkten Tabellenwertebereich für die DLT gehören, aus dem minimalen Tabellenwert (Vmin) und dem maximalen Tabellenwert (Vmax) bestehen, wobei der minimale Tabellenwert (Vmin) und der maximale Tabellenwert (Vmax) unter Verwendung von vorzeichenlosen ganzen Zahlen codiert werden.

8. Verfahren gemäß Anspruch 7, das weiter ein Codieren eines ersten Flags aufweist, das anzeigt, ob die Vollbereichs-DLT oder eine eingeschränkte DLT zu verwenden ist, wobei die eingeschränkte DLT den eingeschränkten Tabellenwertebereich für die DLT von dem minimalen Tabellenwert (Vmin) bis zu dem maximalen Tabellenwert (Vmax) aufweist.

9. Verfahren gemäß Anspruch 7, wobei der minimale Tabellenwert (Vmin) und der maximale Tabellenwert (Vmax) Vorgabewerte verwenden, wenn der minimale Tabellenwert (Vmin) und die maximale Tabellenwertdifferenz nicht codiert sind, und
wobei der minimale Tabellenwert (Vmin) einen ersten Vorgabewert verwendet, der zu -1 korrespondiert, und der maximale Tabellenwert (Vmax) einen zweiten Vorgabewert verwendet, der zu MAX_DEPTH_VALUE + 1 korrespondiert, wobei MAX_DEPTH_VALUE zu dem maximalen Tiefenwert korrespondiert.

10. Verfahren gemäß Anspruch 7, wobei die Bit-Zuordnungswerte für die Tiefenwerte in dem Tiefenbereich unter Verwendung von Lauflängen-Codierung auf Null-Läufen oder Eins-Läufen codiert werden, die zu den Bit-Zuordnungswerten gehören.

11. Verfahren gemäß Anspruch 10, wobei zweite Differenzen zwischen den Null-Läufen und einem minimalen Null-Lauf oder zwischen den Eins-Läufen und einem minimalen Eins-Lauf codiert werden.

12. Verfahren gemäß Anspruch 11, das weiter ein Codieren eines Flags aufweist, das anzeigt, ob die DLT unter Verwendung der Lauflängen-Codierung codiert ist oder die DLT direkt als eine Bit-Zuordnung codiert ist.

13. Verfahren gemäß Anspruch 11, wobei die zweiten Differenzen unter Verwendung von n-Bit-Festlängen-Codierung in die codierten DLT-Informationen codiert werden, wobei n eine positive ganze Zahl ist.

14. Verfahren gemäß Anspruch 13, wobei ein überschüssiger Teil einer aktuellen zweiten Differenz unter Verwendung eines Exp-Golomb-Codes codiert wird, wenn die aktuelle zweite Differenz einen maximalen n-Bit-Wert überschreitet, der durch n Bits repräsentiert wird, wobei der überschüssige Teil zu einer dritten Differenz zwischen der aktuellen zweiten Differenz und dem maximalen n-Bit-Wert korrespondiert.

Revendications

1. Procédé de décodage d'une table de consultation de profondeur, DLT, dans un système de codage vidéo tridimensionnel ou multi-vues, le procédé comprenant le fait :

de recevoir des données d'entrée associées à des informations de la DLT codées d'une carte de profondeur actuelle ;

de décoder une plage de valeurs de table contrainte pour la DLT comprise entre une valeur de table minimale (Vmin) de la DLT et une valeur de table maximale (Vmax) de la DLT sur la base des informations de la DLT codées, et où la valeur de table minimale (Vmin) est supérieure à une valeur de profondeur minimale d'une DLT pleine plage, la valeur de table maximale (Vmax) est inférieure à une valeur de profondeur maximale de la DLT pleine plage, ou la valeur de table minimale (Vmin) est supérieure à la valeur de profondeur minimale et la valeur de table maximale (Vmax) est inférieure à la valeur de profondeur maximale ;

de décoder des valeurs de carte binaire d'une carte binaire (BM) pour des valeurs de profondeur dans une plage de profondeur comprise entre (la valeur de table minimale + 1) et (la valeur de table maximale - 1), où les valeurs de carte binaire indiquent si les valeurs de profondeur sont présentes dans la DLT respectivement ; et

de reconstruire la DLT pour la carte de profondeur actuelle sur la base de la plage de valeurs de table contrainte pour la DLT et des valeurs de carte binaire pour les valeurs de profondeur dans la plage de profondeur ;

dans lequel la plage de valeurs de table contrainte pour la DLT est indiquée par la valeur de table minimale (Vmin) et une différence de valeur de table maximale dans les informations de la DLT codées, où la différence de valeur de table maximale correspond à une première différence entre la valeur de table maximale (Vmax) et la valeur de table minimale (Vmin), et où la valeur de table minimale (Vmin) et la différence de valeur de table maximale sont codées en utilisant des nombres entiers non signés ; ou

dans lequel la plage de valeurs de table contrainte pour la DLT est indiquée par la valeur de table minimale (Vmin) et la valeur de table maximale (Vmax) codées dans les informations de la DLT codées, où la valeur de table minimale (Vmin) et la valeur de table maximale (Vmax) sont codées en utilisant des nombres entiers non signés.

2. Procédé de la revendication 1, comprenant en outre le fait de décoder un premier drapeau pour indiquer s'il faut utiliser la DLT pleine plage ou une DLT contrainte, où la DLT contrainte a la plage de valeurs de table contrainte pour la DLT comprise entre la valeur de table minimale (Vmin) et la valeur de table maximale (Vmax).

3. Procédé de la revendication 1, dans lequel la valeur de table minimale (Vmin) et la valeur de table maximale (Vmax) utilisent des valeurs par défaut lorsque la valeur de table minimale (Vmin) et la différence de valeur de table maximale ne sont pas codées dans les informations de la DLT codées et où la valeur de table minimale (Vmin) utilise une première valeur par défaut correspondant à - 1 et la valeur de table maximale (Vmax) utilise une deuxième valeur par défaut correspondant à MAX_DEPTH_VALUE + 1, où MAX_DEPTH_VALUE correspond à la valeur de profondeur maximale.

4. Procédé de l'une quelconque des revendications précédentes, dans lequel les valeurs de carte binaire pour les valeurs de profondeur dans la plage de profondeur sont codées en utilisant un codage à longueur de série sur des séries de zéro ou des séries de un associées aux valeurs de carte binaire.

5. Procédé de la revendication 4, dans lequel des deuxièmes différences entre les séries de zéro et une série de zéro minimale ou entre les séries de un et une série de un minimale sont codées, ledit procédé comprenant en outre le fait
de décoder un paramètre n à partir des informations de la DLT codées, où les deuxièmes différences sont codées en utilisant un codage à longueur fixe de n bits et n est un nombre entier positif, et
de décoder une partie en excès d'une deuxième différence actuelle à partir des informations de la DLT codées lorsque ledit codage à longueur fixe de n bits indique que la deuxième différence actuelle dépasse une valeur de n bits maximale représentée par n bits, où la partie en excès est codée en utilisant un code Exp-Golomb, et où la partie en excès correspond à une troisième différence entre la deuxième différence actuelle et la valeur de n bits maximale.

6. Procédé de la revendication 4, dans lequel un drapeau est signalé pour indiquer si la DLT est codée en utilisant le codage à longueur de série ou si la DLT est codée sous la frome d'une carte binaire directement.

7. Procédé de codage d'une table de consultation de profondeur, DLT, dans un système de codage vidéo tridimensionnel ou multi-vues, le procédé comprenant le fait :

de recevoir des données d'entrée associées à la DLT d'une carte de profondeur actuelle ;

de déterminer une plage de valeurs de table contrainte pour la DLT comprise entre une valeur de table minimale (Vmin) et une valeur de table maximale (Vmax) de la DLT, où la valeur de table minimale (Vmin) est supérieure à une valeur de profondeur minimale d'une DLT pleine plage, la valeur de table maximale (Vmax) est inférieure à une valeur de profondeur maximale de la DLT pleine plage, ou la valeur de table minimale (Vmin) est supérieure à la valeur de profondeur minimale et la valeur de table maximale (Vmax) est inférieure à la valeur de profondeur maximale ;

de coder des valeurs de carte binaire d'une carte binaire (BM) pour des valeurs de profondeur dans une plage de profondeur comprise entre (la valeur de table minimale + 1) et (la valeur de table maximale - 1), où les valeurs de carte binaire indiquent si les valeurs de profondeur sont présentes dans la DLT respectivement ; et

de coder la DLT en des informations de la DLT codées, où ledit codage de la DLT en informations de la DLT codées comprend le codage de premières informations associées à la plage de valeurs de table contrainte pour la DLT en informations de la DLT codées ;

dans lequel lesdites premières informations associées à la plage de valeurs de table contrainte pour la DLT comprennent la valeur de table minimale (Vmin) et une différence de valeur de table maximale, où la différence de valeur de table maximale correspond à une première différence entre la valeur de table maximale (Vmax) et la valeur de table minimale (Vmin), et où la valeur de table minimale (Vmin) et la différence de valeur de table maximale sont codées en utilisant des nombres entiers non signés ; ou

dans lequel lesdites premières informations associées à la plage de valeurs de table contrainte pour la DLT comprennent la valeur de table minimale (Vmin) et la valeur de table maximale (Vmax), où la valeur de table minimale (Vmin) et la valeur de table maximale (Vmax) sont codées en utilisant des nombres entiers non signés.

8. Procédé de la revendication 7, comprenant en outre le fait de coder un premier drapeau pour indiquer s'il faut utiliser la DLT pleine plage ou une DLT contrainte, où la DLT contrainte a la plage de valeurs de table contrainte pour la DLT comprise entre la valeur de table minimale (Vmin) et la valeur de table maximale (Vmax).

9. Procédé de la revendication 7, dans lequel la valeur de table minimale (Vmin) et la valeur de table maximale (Vmax) utilisent des valeurs par défaut lorsque la valeur de table minimale (Vmin) et la différence de valeur de table maximale ne sont pas codées, et
dans lequel la valeur de table minimale (Vmin) utilise une première valeur par défaut correspondant à - 1 et la valeur de table maximale (Vmax) utilise une deuxième valeur par défaut correspondant à MAX_DEPTH_VALUE + 1, où MAX_DEPTH_VALUE correspond à la valeur de profondeur maximale.

10. Procédé de la revendication 7, dans lequel les valeurs de carte binaire pour les valeurs de profondeur dans la plage de profondeur sont codées en utilisant un codage à longueur de série sur des séries de zéro ou des séries de un associées aux valeurs de carte binaire.

11. Procédé de la revendication 10, dans lequel des deuxièmes différences entre les séries de zéro et une série de zéro minimale ou entre les séries de un et une série de un minimale sont codées.

12. Procédé de la revendication 11, comprenant en outre le fait de coder un drapeau pour indiquer si la DLT est codée en utilisant le codage à longueur de série ou si la DLT est codée sous la frome d'une carte binaire directement.

13. Procédé de la revendication 11, dans lequel les deuxièmes différences sont codées en informations DLT codées en utilisant un codage à longueur fixe de n bits, où n est un nombre entier positif.

14. Procédé de la revendication 13, dans lequel une partie en excès d'une deuxième différence actuelle est codée en utilisant un code Exp-Golomb lorsque la deuxième différence actuelle dépasse une valeur de n bits maximale représentée par n bits, où la partie en excès correspond à une troisième différence entre la deuxième différence actuelle et la valeur de n bits maximale.

Drawing